Publishers are often interested in the conversion of large collections of paper-based documents into digital forms that are suitable for electronic archival purposes and digital libraries. Paper documents may be scanned and converted into digital high-resolution images. These digitial high-resolution images may not be suitable for electronic archival purposes and digital libraries without human intervention to correct and format the images. With a large volume of the material being processed, it becomes increasingly costly to include human beings in the process to correct the automated recognition processes.